Tubelab PCBs for an 845 - a good idea?

[otto88]

Hi

Two questions, re volts and the circuit.

I was considering using tubelab’s Tubelab SE http://tubelab.com/TubelabSE.htm for an 845.
The operating point isn’t chosen yet, but likely will be around 1000 v.

When I mentioned this to a guy familiar with materials (and tubes) he said:

“PCBs for an 845 - NOT a good idea
There's a VERY good reason for this, VOLTs.. lots of 'em (1 KV+) and they can jump/arc across tracks!
Fibreglass PCB substrate ain't a very insulative dielectric at HV, air is much better”
A guy here will be using the Tubelab SE boards for a GM-70.

Are these boards safe in this respect for high volts?

Also, the Tubelab SE has been tried with dozens of tube, mostly 45's, 2A3's and 300B's.

How optimal would the circuit be for an 845 run into A2, with a driver of the PowerDrive, or maybe triode connected EL34, EL36 or 807?

Thanks

 

[kenpeter]

http://tubelab.com/845SE.htm

 

[otto88]

Ken,

That demonstrates that George has done it. I know he’s a very experienced and innovative tube guy

But with up to 1200 volts, it’d be good to know how a problem with arcing on a pcb is avoided.

I also believe him that the powerdrive and A2 improve the sound.

But before committing to a big project like an 845, it’d be good to know *how the circuit which was developed as a generic circuit for other tubes, was optimised for the 845

Cheers

 

[billr]

Hi

the HT for the 845 is not on the PCB. It has 400v for the driver valves.

so it should be ok, as that is within normal parameters.

Hope that helps

Bill

 

[otto88]

Bill,

Thanks, that explains the volts question.

Hopefully George (who I emailed re this thread) will chime in on how the circuit was or can optimised for the 845.

Cheers

 

[nhuwar]

You can still put 1.2 kv on a pcb as long as it's a glass epoxy pcb. If you have I'd say .125" gap between the + &- traces you shold be fine. Remember 30kv per centimeter in air is the instantaneous breakdown distance.


Look for a pcb made from g10 and you will never have a problem even up to 6 or 7 kv.








Nick

 

[Tubelab_com]

Also, the Tubelab SE has been tried with dozens of tube, mostly 45's, 2A3's and 300B's. ..... or maybe triode connected EL34, EL36 or 807?

The Tubelab SE board is a two stage amplifier PCB. It was optimized for use with 45's, 300B's and 2A3's (in that order). The board is designed for mounting a 4 pin socket that fits these DHT's directly on the PC board. I have used it with other tubes for testing purposes. Serious modification is needed to use a tube that does not have a 4 pin base in the Tubelab SE board.

But before committing to a big project like an 845, it’d be good to know *how the circuit which was developed as a generic circuit for other tubes, was optimised for the 845

The Tubelab SE isn't a generic board it was designed for use with the 3 tube types mentioned above. An 845 requires a bunch of grid drive (several hundred volts P-P). It is highly unlikely that a succesful 2 stage 845 amp could be built. What I did with my 845 amp was to use the Tubelab SE as the first two stages. In this case it functions pretty much as a normal Tubelab SE running 45 tubes. The quiescent plate voltage on the 45 is about 250 to 275 volts. I used a "PowerDrive" board to interface the Tubelab SE to the 845, creating a 3 stage amp. At least one builder has used a Tubelab SE to drive an 845 via an interstage transformer.

Some information on the PowerDrive circuit can be found here:

http://www.tubelab.com/powerdrive.htm

Are these boards safe in this respect for high volts? .....

No they are not. The Tubelab SE was designed for peak plate voltages in the 500 volt range. In my 845 amp, no extreme voltages are connected to the Tubelab SE PC board. The HV is only connected to the primary of the OPT which is connected to the plate of the 845. The peak plate voltage on the 845 can go over 2000 volts. The power supply and the OPT wiring are exposed to high voltages and special construction practices and components are required for use at these voltage levels.

There's a VERY good reason for this, VOLTs.. lots of 'em (1 KV+) and they can jump/arc across tracks! .... But with up to 1200 volts, it’d be good to know how a problem with arcing on a pcb is avoided.

I did use a PC board in my HV power supply which puts out 1100 volts. The board must be designed for operation at extreme voltages. This is extremely important in humid climates (like mine), since moisture, dust, and high voltages can lead to disaster. Accidental contact with the voltages in this supply could result in DEATH! This must be carefully considered during the design of any 845 amp. All of the possible failure modes must be planned for.

How optimal would the circuit be for an 845 run into A2, with a driver of the PowerDrive, or maybe triode connected EL34, EL36 or 807?

A properly implemented PowerDrive circuit can drive an 845 or an 833A well into A2 without distortion without consuming power (drawing current) from the driver tube. I chose the 45 tube for my driver because of its great linearity.

A well designed transformer coupled design could be used in moderate A2. The driver tube must be capable of supplying power gince the grid current comes from the driver tube.

I put the schematics for the PowerDrive and for my 845SE amp on my web site for EXPERIENCED builders to use as a point of reference for their amp designs. An 845 amp is not for inexperienced constructors. The parts are expensive since many of them must be rated for multi kilovolt operation (transformers, capacitors, and even the wire). The probability of failure is much higher than a simple amp due to the circuit complexity and the extreme voltages. The price of failure is much higher since it is possible to fry some expensive parts, or worse.

 

[otto88]

George,

Based on your case study described at the PowerDrive cookbook page, I also intend to run into A2, primarily for the lower distortion. (Though I’m also after 35-40 watts).

A friend who’s more tube knowledgeable than me, who would be doing the more critical parts of the project, said re A2:

“The main constraint on the driver is undistorted voltage output at 100's Vp-p and the ability to supply a few milliamps of current at the +ve peak without distorting. This may eliminate the 45”

I’m also attracted to using the 45 as driver (if not, possibly a triode strapped EL34, EL36 or 807 – meaning not Tubelab SE boards).
With appropriate PowerDrive (rather than transformer coupled) have you found or do you see any 'limitations' using the 45 as the driver for the 845 into A2?

Thanks

 

[otto88]

I just re-read the article on the 845 SE amplifier itself, which I hadn’t read for a while.

The best version of the amp uses mosfets both after *and before* the 45 driver stage.

So apparently that overcomes any limitations using the 45 as the driver for the 845 into A2?

 

[Tubelab_com]

The PowerDrive circuit seperates the voltage amplification and the current sourcing functions. The tube (45 in this case) provides the voltage gain. The CCS and the mosfet follower allows the tube to operate with a very high load impedance so it does not have to provide any drive current even in A2. The tube provides pure voltage amplification. The 45 is very well suited for this task.

The mosfet buffer sources all of the drive current for the output tube. I took the unmodified 845 am and wired it up to an 833A tube. The Power Drive circuit had no problems with grid current peaks of several hundred mA. Power output was in the 200 watt range.

http://www.tubelab.com/833SE.htm

 

[crazyfrog]

The Tubelab SE board is a two stage amplifier PCB. It was optimized for use with 45's, 300B's and 2A3's (in that order). The board is designed for mounting a 4 pin socket that fits these DHT's directly on the PC board. I have used it with other tubes for testing purposes. Serious modification is needed to use a tube that does not have a 4 pin base in the Tubelab SE board.



The Tubelab SE isn't a generic board it was designed for use with the 3 tube types mentioned above. An 845 requires a bunch of grid drive (several hundred volts P-P). It is highly unlikely that a succesful 2 stage 845 amp could be built. What I did with my 845 amp was to use the Tubelab SE as the first two stages. In this case it functions pretty much as a normal Tubelab SE running 45 tubes. The quiescent plate voltage on the 45 is about 250 to 275 volts. I used a "PowerDrive" board to interface the Tubelab SE to the 845, creating a 3 stage amp. At least one builder has used a Tubelab SE to drive an 845 via an interstage transformer.
Some information on the PowerDrive circuit can be found here:

PowerDrive

.

Who? What transformer? That's what I want to do with a 211. Details please!

 

[dhaen]

PCB's for high voltage use often use slots around the HV tracks. This avoids to some extent, the problem of dust build-up caused by static attraction, which could otherwise cause arcing or discharge.

 

[Tubelab_com]

PCB's for high voltage use often use slots around the HV tracks

I wouldn't run 1100 volts through a PCB unless adequate spacing (no excessive spacing) was used. The Tubelab SE board used in my 845SE sees no more than 400 volts. The power supply board does develop the 1100 volt B+ but the spacing between ground and the 1100 volt line is about 2 INCHES. The amp has been in existence for about 6 years in the south Florida humidity with no issues but I don't use it much because it dumps about 500 watts of heat into the room. Slots are used in some older HV boards but it isn't always a good idea unless the boards are properly coated after milling the slots. The fiberglass used in the laminate can be slightly hydroscopic and it can absorb enough moisture if exposed.

Who? What transformer? That's what I want to do with a 211. Details please!

People who build my amps often send me pictures of the finished units. I got a picture of an 845 SE amp from a builder in Europe several years ago. His discription stated that it was transformer driven, but no details were given.

I have never been a big fan of driver transformers since I believe that a mosfet follower is a better solution, but there are some nice sounding amps out there that use them. Both the 845 and the 211 respond well to A2 operation. The 211 is the easier of the two to drive and a 45 can develop the voltage to drive it easilly. Whether it can push enough current into the grid of a 211 to squeeze 40+ watts out of it without a little help from some sand, I don't know. Even if you don't expect to fully saturate the 211 you should be able to get 30 to 35 watts using a 45 for a driver and a 1:1 driver transformer.

I would look for a transformer that goes from 5K ohms to 5K ohms that can eat 25 to 30 mA of primary current (the sweet spot for a 45). You will want a good resistive load on the secondary in the 5K to 10 K range. Measure the frequency response at the driver transformer secondary with no 211 in the socket and no high voltge turned on. Tweak the resistor value to get a flat response with minimum distortion. A 211 will need about 150 volts P-P of clean drive, a bit more if you want to do A2. Your negative bias should be able of going to -80 volts or so.

If your negative bias can go to -200 volts and you can get 300 to 350 volts P-P of drive, you can also run 845's in the same amp. I can swap between the two tubes by readjusting the bias in my amp.

 

[crazyfrog]

I wouldn't run 1100 volts through a PCB unless adequate spacing (no excessive spacing) was used. The Tubelab SE board used in my 845SE sees no more than 400 volts. The power supply board does develop the 1100 volt B+ but the spacing between ground and the 1100 volt line is about 2 INCHES. The amp has been in existence for about 6 years in the south Florida humidity with no issues but I don't use it much because it dumps about 500 watts of heat into the room. Slots are used in some older HV boards but it isn't always a good idea unless the boards are properly coated after milling the slots. The fiberglass used in the laminate can be slightly hydroscopic and it can absorb enough moisture if exposed.



People who build my amps often send me pictures of the finished units. I got a picture of an 845 SE amp from a builder in Europe several years ago. His discription stated that it was transformer driven, but no details were given.

I have never been a big fan of driver transformers since I believe that a mosfet follower is a better solution, but there are some nice sounding amps out there that use them. Both the 845 and the 211 respond well to A2 operation. The 211 is the easier of the two to drive and a 45 can develop the voltage to drive it easilly. Whether it can push enough current into the grid of a 211 to squeeze 40+ watts out of it without a little help from some sand, I don't know. Even if you don't expect to fully saturate the 211 you should be able to get 30 to 35 watts using a 45 for a driver and a 1:1 driver transformer.

I would look for a transformer that goes from 5K ohms to 5K ohms that can eat 25 to 30 mA of primary current (the sweet spot for a 45). You will want a good resistive load on the secondary in the 5K to 10 K range. Measure the frequency response at the driver transformer secondary with no 211 in the socket and no high voltge turned on. Tweak the resistor value to get a flat response with minimum distortion. A 211 will need about 150 volts P-P of clean drive, a bit more if you want to do A2. Your negative bias should be able of going to -80 volts or so.

If your negative bias can go to -200 volts and you can get 300 to 350 volts P-P of drive, you can also run 845's in the same amp. I can swap between the two tubes by readjusting the bias in my amp.

Ahh! Thank you Georges, that information will be very helpful!
(BTW I got my Tubelab SE board today! )

 

[tim845]

I've "stolen" Georges's concept without his permission to drive my friend's two stages 211SE into a 30W sinewave unclipped power output at an 8 ohms dummy load

The config. was very simple before:

Paralleled 6463 in text book common cathode circuit but loaded by a 1:1 interstage then coupled to 211, 211 is at self bias.

My friend is using expensive italy made interstage and output transformer but still got non-satisfactory sound and power.

We also tried 5687 as cathode follower to drive 211 before and sound improved but same result - "no power"

Later, based on Tubelab's concept, we tried to use IRF840 wired as soure follower to drive 211, now the power is "ON" !!!

Now the source follower is replaced by 2SK2700 because of its Crss is lower than IRF840.

The following measurement of the 211SE of my friend is as follows (for reference only)

1) +/-200VDC to feed 2SK2700 and adjust the output of the source follower can feed the grid of 211 upto -45VDC

2) the Vgs of 2SK2700 is +3.19V approimately

3) the source resistor of 2SK2700 is 5K 20W (the former was 10K). The total current drawn by the follower is 25ma approximately. Vds is 245VDC.

4) Tube 211 is Made in China, the driver tube is 6463 in parallel connection.

5) 960VDC feeds to the plate of 211

BTW, after many times of serious listening by many "serious friends", no one knows why 211 can fire this power but just find themselves an excuse of the excellent quality of the interstage and output transformer lead to this result.

No one claims that there is any silicon sound !!! Because no one knows a MOSFET source follower built inside

Thanks Georges

 

[Steven-H]

Even if you don't expect to fully saturate the 211 you should be able to get 30 to 35 watts using a 45 for a driver and a 1:1 driver transformer.

I would look for a transformer that goes from 5K ohms to 5K ohms that can eat 25 to 30 mA of primary current (the sweet spot for a 45). You will want a good resistive load on the secondary in the 5K to 10 K range. Measure the frequency response at the driver transformer secondary with no 211 in the socket and no high voltge turned on. Tweak the resistor value to get a flat response with minimum distortion. A 211 will need about 150 volts P-P of clean drive, a bit more if you want to do A2. Your negative bias should be able of going to -80 volts or so.

This is exactly my plan; but to drive an 813. I already have the transformers from Electraprint...Now i just need to find the time to build the damn thing